This week, Emma gained access to a sewing machine and was able to complete the headband (pending further modifications). The fabric chosen for the band sleeve (a Spandex-polyester mix) is a stretchy material that is somewhat delicate, so she encountered some problems with it getting sucked into the sewing machine. Additionally, the elastic thread chosen for use with the sleeve construction was difficult to use in the sewing machine, so that should be modified for future machine use. Eventually, by backing the seam with felt, the sleeve was sewn. The headband was assembled, and the IMU was inserted into the pocket of the band to be used for calibration.
Emma has also been working on making formal wiring diagrams for the final report and going through the DesignSafe analysis. This week, Katelyn struggled with getting the correct input from an Arduino to the Pi. Dr. Klaesner showed her that the issue was that the Serial Monitor for the Arduino was open (so Arduino IDE was trying to read data from Serial) at the same time that the Python program on the Pi was trying to read data from Serial. This caused all sorts of distortions and odd effects. With this resolved, Katelyn was able to get the Pi to accurately read output from an Arduino test program that gave values of 0, 1, or 2 every second. She also finished working through the other relevant issues for the Pi - the Python program can now activate audio output and there is a button to halt the program and write a file containing statistics from the session. Thom worked on calibration, and after discussion with Dr. Klaesner, was able to get a calibration loop up and running. He was also able to add cutoff values to the program so that the output is given as a 0, 1, or 2 to match the inputs needed for Katelyn’s Python program that will read in input from the Arduino and give corresponding feedback. Katelyn and Thom have largely completed the device. Individually, they made large strides towards the completed device, and spent Thursday and Friday morning overcoming the final hurdle of getting the Arduino and the Pi to communicate with each other so the verification and validation can be completed this weekend. However, it is still difficult to get the device to work consistently - it is proving very difficult to pinpoint the issue that is causing problems with the Arduino, and the devices worked together once, then immediately stopped working again. We have had problems with the Arduino running for a short period of time, then stopping, and are not sure if the Pi is causing this problem. To solve this, we have been running tests using the Arduino with our personal computers - it does not appear to be a problem with the Arduino, so it may be an issue with running the program on the Pi. We also need to start being more consistent about the order in which we run the Arduino and Python code, since this seems to have an effect. Our goal is to have it working more consistently in order to be able to test it well. Over the next week, we will be completing our verification and validation and writing our final report. Emma is working to finalize the physical headband. The base of the headband has been constructed, and she has verified that it fits our device specifications in terms of the amount of the head that it covers. She is working to secure the use of a sewing machine in order to construct the pocket that will hold the IMU in the headband, thereby completing the headband portion of the device. Emma also began looking into the final report and particularly the DesignSafe analysis. The DesignSafe program proved to be confusing and potentially difficult to use with our product, but she is currently gathering more information about how we can best use it for our device.
Thom got the calibration loop working this week - a button press activates a loop in the Arduino program. However, it triggers some sort of FIFO (first in first out) overflow that causes NaN (not a number) to be displayed for one of our yaw-pitch-roll values. He is currently working to resolve this issue. Katelyn is working with the Raspberry Pi and got some advice from a friend about how best to avoid timing issues when getting input from the Raspberry Pi from the Arduino Serial monitor. However, she is getting input from the Serial monitor, and just needs to figure out how to fine tune the timing (some of the inputs get cut off partway) and how to get the correct output from the Pi as a result. We are continuing to work on developing our prototype. Emma is developing our final version of the headband, using Thom’s head as a size model and working to make a design that would be fairly easily adapted to any size head. However, for the purposes of this project, we are only planning to make the headband in one size. It will contain a pocket for the IMU, which Emma plans to make relatively snug to keep the IMU from shifting as much as it may in our current version of the headband, which is much less precise of a prototype (the pocket is bigger and looser, and the headband is not adjustable.
Katelyn is continuing to work with the Raspberry Pi, and hopes to have the audio/haptics working based on input from the Arduino Serial monitor by early next week. This may or may not be achievable, but that mostly depends on how smoothly getting input from the Serial monitor goes. Once she has achieved these tasks, she plans to switch to helping Thom get the IMU measurement/calibration improved, so that the IMU can be hooked up with the Raspberry Pi to both measure head position and give feedback. Her other upcoming task will involve modifying the Python program on the Pi to save data over time in a format that would be helpful for the clinician. Thom is continuing to try to get the Arduino to have a calibration component to its program, activated by a button press, in order to normalize the IMU values. This week, we began to work in earnest again on our prototype. Thom is working on building a calibration system into the Arduino program. He is working on adding a button to the Arduino that will trigger a calibration process when pressed. Initially, this button seemed to interfere with the output of the IMU. We will investigate the circuits built to try to resolve this issue, so that we will be able to use this button without triggering IMU problems. Thom will also be working to write the code that will allow the IMU to be calibrated for its position on a user’s head.
Emma drew up a plan for the rest of the semester, outlining what tasks have to be completed in what order. She also began working on the headband itself. She obtained fabrics from Joann’s, including felt, swimsuit liner, and elastic, and is working to construct a safe, comfortable headband that will securely hold the IMU in place. She will continue construction of the headband into next week. Katelyn is working with the Raspberry Pi and the haptic motor. She wrote a short Python program that causes the haptic motor to vibrate (with the haptic motor connected to a GPIO pin and to ground). This required Emma to solder the haptic motor to two longer wires that connect well to a breadboard. For next week, she is researching the syntax to use in Python in order to get data from the Arduino that could determine the output from the haptic motor. She will also learn how to write code in Python that will start and stop audio output. This week, our focus was the Verification and Validation presentation. Over the weekend, Emma prepared her presentation, and we went over it with Harry, our TA. He primarily requested for more clarity and detail about which parts of our verification and validation we will actually be able to do within the scope of this course. He also wanted us to be more detailed with how we plan to verify that we are meeting our goals, such as “safe” and “accurate.”
Emma gave a great presentation on Monday. Also, Katelyn ordered a small haptic motor to be operated via the Raspberry Pi for haptic feedback. We will resume active work on our prototype after spring break. This week, our focus was on the Verification and Validation report, which is due on Friday. Emma wrote the Verification and Validation sections, with input and feedback from Katelyn and Thom. Katelyn wrote the Project Status and Proof-of-Concept Testing section, with editing from Thom on the parts of the project that he knows best. Emma also updated the specs for the device and wrote surveys that we will eventually use for verification and validation.
Thom continues to work on the IMU. We met with Dr. Klaesner on Monday to get some help with understanding the data being provided by the IMU. Specifically, we discussed calibration and potentially including a calibration loop controlled by a button that would zero out the values for pitch, yaw, and roll. We also discussed how to put the whole program together using the Raspberry Pi to get data from the IMU/Arduino unit, process it, and return feedback to the user. The obstacle that we are currently facing is getting the IMU data to the Raspberry Pi in real time. We may download a Raspberry Pi emulator to do some testing, as per Dr. Klaesner’s suggestion. Emma will be working on the V&V presentation this weekend, but we hope to shift back to working full-time on the prototype as soon as we can. Thom continues to work on the IMU. His current thought is that the offsets don’t need to be too specific because they change with initial head position every time our Arduino program is run. It will be more important to account for the initial head position. He also investigated which dimension (pitch, yaw, or roll) to use for each direction the head might move in (forward, right, left, back). In order to investigate this, he is taking the data (pitch, yaw, and roll measured over time) outputted by the IMU from the Arduino serial output and plotting it in MATLAB. In order to plot this accurately, he also investigated how often the IMU readings are reported.
In order to facilitate the testing of the IMU, Thom is investigating ways to interact with Arduino from MATLAB. Ideally, we could get MATLAB to automatically plot the data from the Arduino serial output without manual input. We also met with Dr. Klaesner in order to ask some questions. We spoke about the calibration of the IMU, which Dr. Klaesner suggested we may have to do for each individual user by taking several points of reference before beginning the feedback cycle. We also asked about pulling Arduino data into MATLAB, and Dr. Klaesner suggested investigating exactly what data is being provided by the IMU and how the Arduino program we have been using processes this data to produce values for pitch, yaw, and roll. Thom is looking into what data is going into each pin of the Arduino, and we will meet with Dr. Klaesner in the lab next Monday to examine this subject with him. Katelyn is researching the best way to interface the Arduino and the Raspberry Pi. In particular, it seems like most programming on the Raspberry Pi is done in Python, so she is going to be looking at how the Arduino IDE and a Python program could potentially be used in conjunction in order to activate feedback based off the results provided by the Arduino/IMU. She is also planning to order haptic motors this week, so that we can test them with the Pi and/or the Arduino next week or the week after. Emma is working on writing our V+V report. We had some questions about how the testing can be performed for certain areas for Dr. Klaesner this week, since it will be impossible to perform some forms of testing with our prototype device, the only one of its kind. We hope to finish this report quickly to be able to proofread it well and have time to continue working on our prototype next week. As usual, Thom is working on programming the IMU. The IMU program that we are using (found here: https://maker.pro/arduino/tutorial/how-to-interface-arduino-and-the-mpu-6050-sensor) accepts X, Y, and Z offsets, so Thom is working to determine how these offsets affect the measured pitch, yaw, and roll. He plotted the Arduino output in MATLAB, but the next step is to communicate between the two programs and get continuous processing. This will be useful for determining the correct offsets and for deciding how to process the data.
The ability to measure pitch, yaw, and roll will affect the placement of the IMU on the head. We currently are planning to position the IMU on the side of the head, since it is difficult for people to wear headbands such that part of the headband is flat on top of the head. Although we are not currently sure how to make this happen, we would like to have some sort of zero button to reset the IMU position. Ideally, this zero button would help compensate for slightly different positioning of the IMU on each person’s head and for different levels of ability to raise the head fully. Emma is currently working on the V+V report, and has been going through our list of device specifications to determine how we could potentially test each item. As she goes through the list of specs, she is keeping track of questions that come up. We plan to meet with Dr. Klaesner within the next week in order to get our questions about the V+V report cleared up. Katelyn contacted our client, Dr. Courtney Dunn, with updates on how our device is going and with some questions. The questions and answers are logged in LabArchives. She also is looking into the use of haptic motors with Raspberry Pis and has been taking care of moving our Google Drive notes into LabArchives and other small tasks. Thom is continuing to work on programming the IMU to accurately detect position. He has familiarized himself with the IMU code and how the program we downloaded to calculate IMU position works. He has modified the program to use the value for pitch to give a binary value for head up or head down. However, we still have a lot of work to do in order to make this evaluation of “up” or “down” more precise and usable in an accurate setting, which is what Thom will focus on continuing to do.
We acquired a headband and some fabric, and Emma sewed a pocket into the headband to hold the IMU. Thom will be working to find the best position for the IMU in the headband - how best to wear the headband in order to get accurate IMU measurements. We received our new monitor for the Pi, and are now using it exclusively. On Monday, Emma and Katelyn experimented with the Bluetooth and audio capabilities of the Raspberry Pi, and successfully managed to get the Pi to play a downloaded .mp3 file through VLC player through the HDMI monitor, through the 3.5 mm headphone jack, and through Bluetooth earbuds. Katelyn’s next steps are to look at haptic motors that work with the Raspberry Pi and order haptic motor parts so we can try to get the Pi to control output. After that, she will be working to write code on the Pi to coordinate the Pi’s output of audio and haptics. We are starting to work on our V+V report, and hope to have most of it written in the next week or two. We may have to change our meeting schedule somewhat, as Katelyn will be out of town for the next four weekends. As we began assembly of our prototype this week, we encountered several problems of varying difficulties to address. Our primary problem right now, though solvable, is a lack of monitor into which we can plug our Raspberry Pi unit for programming. Though we tried several workarounds, we currently have no way of interfacing with the unit other than a television in one of our homes. Since this is infeasible for frequent work, we have purchased a small monitor for use with the Pi. Data has been successfully transmitted from the IMU, but not on a consistent basis. We have had some trouble with our IMU code cutting out; we soldered on the IMU pins today, and hope this solves any issues related to connectivity.
Next week, due to our difficulties last week, our goals will resemble those of last week: achieve successful data transmission from the IMU to the Arduino, develop initial programming for the Pi, and connect all the parts we can. While we wait several days for a portable monitor, a TV in someone’s home can be used. Due to some malfunctioning programming originally intended to help project the Pi screen, we may have to wipe the card and start from the factory settings. This is fortunately early in the process and there is not much on the card, but it will be a small setback if necessary. We are also looking into methods of backing up the Pi’s SD card so that we can have restore it to a working state if necessary. This week, most of our parts arrived that we will need for initial assembly of the bare-bones device. Most importantly, we received the IMU and Raspberry Pi, which we should be able to use along with an Arduino unit and SD card to start work. Once these arrived, we were able to learn more about the components necessary to operate the Raspberry Pi unit - most of these we have or can easily acquire, which we will be doing over the next week. It also required a change of settings on the microSD card to the NOOBS operating system; this change was accomplished on Thursday, and should allow the memory card to be compatible with the Pi. We will also have to install Rasbian on the Pi unit.
Over the weekend, we all will meet to begin assembly of the parts we have. To prepare, we will begin reading through manuals for the parts, since it is important that we are all able to operate the system. Emma will locate guides for the IMU since it does not come with instructions; she will then become the team ‘expert’ on this material. Based on how the weekend’s meeting goes, we will plan our work for the upcoming week. Our goals for next week include connecting the IMU to the Arduino such that data is transmitted, developing initial programming for the Pi, and having all available parts correctly connected. This week, we ordered a Raspberry Pi 3 B+ because it is cost-effective and has 4 USB ports. We also ordered a 32GB SanDisk Extreme microSDHC card for Pi storage because it has wear leveling and will be compatible with the Pi. Both will arrive this Tuesday, January 22nd. Katelyn is looking for a haptic motor that has a USB input and can work with the Pi. Thom already owns a SparkFun RedBoard, so when our components arrive next week we can begin trying to put them together in Professor Widder’s lab.
On Tuesday we met with Professor Klaesner to talk about our progress. We asked about the possibility of using a tilt sensor instead of an IMU, and he pointed us to Maria Mihaelescu, who experimented with both during her project. After consulting with her, we determined that the MPU6050 sensor, a 6-axis IMU, would be best suited to our needs. It is well-tested and extensive documentation is available online, which will be a useful resource for our group since code is one of our weaker points of experience. This should be arriving in the mail today, and we will be able to begin testing and working with the sensor immediately. We also set weekly meeting times that line up with times where Professor Widder’s lab is available. For this week and part of next week, these will be used for planning and ordering; in the future, they will be used for building and testing the device, as well as for coding and getting help from professors. This week, we finished writing and editing the progress report. As before, Katelyn focused on feedback mechanisms, Emma focused on sensors, and Thom focused on methods of computation and transmission of data. It was printed and submitted on Friday morning. The majority of this work was in compiling research, describing the final device, updating the scope and specifications from the preliminary report, and formatting the paper.
In the upcoming week, we will prepare a presentation which Katelyn will deliver on Wednesday. We also plan to be in communication with people from Special School District that Dr. Klaesner has put us in contact with who have expertise on children with CP. We will also begin the process of narrowing down specific components so we can order them over break. This week, we continued to reach out to others to better understand the problem. Katelyn had success in a Facebook group for pediatric physical therapy, finding several people willing to talk about their experiences with head drop, as well as with the CTO of Opter, who offered some information regarding the measurement process that their posture device used. Katie Lammers, a physical therapist, was also contacted regarding the effectiveness of different forms of biofeedback as well as general day-to-day problems that people with CP may face. We are also now in communication with Melissa Reddington, an OT in Francis Howell school district. Hopefully the feedback we get from these people will help us construct our Pugh chart for biofeedback types.
We also continued doing research on the specs of different sensors in order to make our Pugh chart for that component of our project. We are dividing our Pugh chart into several parts - the sensor, the feedback type, the component that will analyze and export the information from the sensor, and the overall device that will contain the other components. We are planning to begin writing our paper this weekend, with the goal of having it mostly written by Thanksgiving. This week, Emma visited Children’s to shadow Courtney as she went between the neurology department and cerebral palsy clinic to examine young children who may be experiencing various developmental delays. Unfortunately from an educational perspective, there were not many consults needed and none were in our target demographic. Emma did, however, see several 1-3 year old children showing early signs of atypical development, which provided some context as to how things may present early on and how other parts of the body may be affected by imbalances in muscle tone. Listening to parents talk was also helpful: it seemed that in many cases, regular access to doctor visits was not possible due to a variety of factors and they were reliant largely upon services provided by Missouri’s First Steps program, which works with schools to provide PT and OT services. This further underscores the need for a device that can be used easily at home without excessive parental interference, since First Steps typically only provides services 1-2 times weekly and parents seemed to cite time and difficulty as reasons that home therapy did not occur as often as recommended.
Our group met several times this week to brainstorm and work through project logistics, as well as to have dedicated work sessions. As of now, our solution is broken down into sensor mechanism and feedback mechanism: this will likely evolve into a progressive Pugh chart with several possible solution styles for the best choices of these two categories, as the rest of the device will depend largely on what we select to fulfill these two specific needs. Next week, we intend to start writing our paper and making concrete process on a Pugh chart to narrow down the set of solutions, allowing us to begin thinking about other design components. During this process, we hope to delve more into our designated areas of research as the project begins to have more discernible sections. We also have concrete plans to continue interacting with healthcare professionals in order to obtain feedback on which of our design directions will most likely be useful in a practical setting. We have joined a Facebook group for St. Louis Pediatric Therapy, and have asked Dr. Klaesner to put us in contact with the Special School District of St. Louis County. This week, Katelyn visited Therapy Services at the Children’s Hospital, like Thom did last week. Her experience was with patients getting checkups - seeing a PT, an orthotist, and a PA to get a general overview of their health and to make adjustments to their treatment plans. While neither patient that she saw was a good candidate for the device we are planning to build, she learned a lot about how CP patients are treated and the many factors that influence their health. In particular, she spoke to medical professionals about the routine care that many CP patients receive. One of the patients she saw attends a school designed for students who have CP and other disabilities. We are hoping to speak with therapy programs in schools to understand what our patients’ capabilities are and the kind of treatments they currently do on a regular basis. Understanding the needs of patients and the environment they are usually in will be critical for designing a device that will be useful.
Because we have been struggling a bit with managing our time, we made a concrete plan to distribute research and work. We set an open-ended time to brainstorm for this upcoming Sunday, during which time we plan to have a dedicated research, brainstorm, and work session to give us a base of knowledge from which to move forward and some potential solutions to investigate. After narrowing our brainstorming list somewhat, we will go through these solutions with Courtney as well as any other PTs/OTs who work with the target population. In order to best understand the problem, we also are compiling a list of groups and people to reach out to for information. In the next week, we will begin to contact them and, if necessary, set up times to meet and/or talk with them about their experiences. These groups and people include Beth O’Neal, First Steps, posture support device companies, special education schools, and families of children with CP. We also will be touching base with Courtney soon after we develop preliminary results of brainstorming. This week, Thom visited Therapy Services at the Children’s Hospital, and had the chance to observe two patients with movement problems. He shadowed during wheelchair consultations and had the opportunity to learn more about how wheelchair components are selected and formed for patients, as well as the types of movement difficulties that patients deal with. We plan to follow up with one of the physical therapists, Beth O’Neal, about other chances to observe patients, specifically patients with movement problems that are directly relevant to our goals. We also plan to shadow during physical therapy sessions sometime in the next two weeks in order to better understand current treatments and how we could develop a device to contribute to physical therapy outcomes at home.
We also continued working on brainstorming this week. We plan to develop several Pugh charts in order to account for the different components of the project we need to solve - obtaining data on head position, providing feedback to the patient, and reporting data to clinicians. Our website, which was submitted on Sunday, will continue to be updated as we progress. This week, we began brainstorming potential solutions to the problem of head drop in patients who have conditions like CP and MS and who use wheelchairs. We are looking at two broad categories of solutions, neck support and biofeedback, and beginning the process of further subdividing categories of solutions.
We are also working to arrange a time with our client, Dr. Courtney Dunn, to see her patients and gain a better understanding of what solutions they are currently utilizing and what they would like to see in a future solution. In particular, we would like to observe firsthand the skill set of the patients to better understand which of our potential solutions would work best for them. This week, Thom gave our Preliminary Presentation and we watched the Preliminary Presentations of the other groups. From watching the other groups, we were able to get some ideas about other design specifications that we might want to consider.
We have begun the process of brainstorming possible solutions that would fulfill our project need/scope. Although this weekend will be Fall Break, we plan to meet or group call at least once to continue developing potential solutions. This will likely involve some research into available sensors and materials, since we will soon need a more concrete picture of each solution to compare the potential solutions based on cost, safety, and other factors. This week, we continued researching existing solutions, especially branching out into patents and commercially available devices. We discussed existing solutions with our client, Dr. Dunn, and Thom reached out to Dr. Kerri Morgan at the Enabling Mobility in the Community Lab, part of WUSTL OT.
We were also focused on writing our Preliminary Report and preparing our Preliminary Presentation this week. This included planning our basic schedule for the rest of the year using the Gantt Chart and dividing out responsibilities. Thom will give our Preliminary Presentation next week, Katelyn will give the Progress Presentation in December, and Emma will give the Verification and Validation Presentation in March. Next week, we will use our knowledge of the existing solution space as we think of potential solutions for head drop. We will likely also continue with research as needed to understand how existing solutions work and how they may be able to be improved upon. This week, we made edits to our need statement and scope to more concisely meet the requirements and describe the problem at hand. We made only one minor change to our design specifications, but plan to come up with more quantitative metrics as the project develops, particularly to quantify comfort. We have continued researching existing solutions, particularly regarding biofeedback. Thom and Emma met to lay out the first presentation and paper with a rough page count and description of content, as well as to talk about website design and content. We also divided up specific roles for writing the next report. Finally, we made a rough schedule for the three presentations. Next week, we plan to complete our first paper, doing research as necessary to supplement our existing knowledge of the subjects discussed
We started out this week by meeting with Dr. Klaesner on Friday 9/14 to discuss possible directions to take with our project, as well as how to define a need statement and scope. As a result of this meeting, that same day Emma and Katelyn were able to visit the Paraquad facility to inspect current wheelchairs and headrests. This visit was also helpful in developing some design specifications, such as the need for a washable cover and a lack of pressure points, and learning more about potential problems that may arise, such as ulcers or sores on the head.
Throughout the week, we did individual research on existing solutions which we are in the process of compiling. Emma focused on creating the website, which is close to being put online (URL will be www.abetterheadrest.weebly.com). As a result of these meetings, we were able to narrow our problem enough to refine our need statement, scope, and initial specifications, which we prepared for turn-in on Monday 9/17. Since edits revealed that these require further work, we plan to revise the need statement and scope for resubmission on Monday 9/24. Particularly, we plan to shorten the need statement to one sentence, remove excess detail from the scope, and specify that we will be working with Dr. Courtney Dunn. In addition to these edits, we plan to continue next week with research on the topics of muscle tone in individuals with cerebral palsy, problems that people using wheelchairs may encounter, and existing biofeedback treatments for head drop. In particular, we plan to research metrics by which the success of existing headrests are measured. The website should be online by the end of next week. This week, we met with Dr. Courtney Dunn, our client, to discuss specifics of two problems she is interested in solving. We had previously discussed the problem of uneven muscle tone and torticollis in infants and potential solutions involving tracking arm movement or head tilt in infants and reinforcing movement of the weaker side through biofeedback. We decided against working with arm weakness due to the success of a previous senior design group with treating the uneven arm tone problem. During our conversation, we discussed existing solutions, which mostly consist of physical therapy, and varying causes of the head tilt.
We also discussed wheelchair headrests that either do not support upright head position or restrain the head uncomfortably or obtrusively. Head drop and head tilt are problems for many people who have cerebral palsy, especially those who use a wheelchair. We decided to pursue a solution to this problem rather than infant torticollis because we think it addresses a need that is not currently being met. Torticollis and head tilt can be effectively treated by physical therapy and surgery already, although the treatment is imperfect and doesn’t work as well for patients with visual impairments. However, wheelchair headrests currently on the market either fail to provide adequate support or are excessively intrusive such that customers are unwilling to use them. We have written a need statement, a project scope, and a list of initial specifications for the head drop problem, and will be revising these to turn in on Monday. Next week, we plan to begin research specific to the problem of muscle tone in people with cerebral palsy. This will include a review of literature pertaining to the problem itself, as well as a preliminary search of existing treatments. After this initial search, we will begin to elaborate upon our current list of specifications to better understand the criteria our solution must meet. We also plan to start on the website if time allows. |
Authors
Thom Ellison, Emma Huff, Katelyn Miyasaki Archives
April 2019
Categories
|